Untitled

Introduction

Background, what’s out there (visualization tools,) why this is useful (because there are not that many detailed examples showing the code, talk about your experience in Sunbelt “what’s the format of the data”, look for papers talking about computing literacy) and our goal (start to finish network visualization: load the data, process it a little bit, and plot it).

Network visualization in a nutshell: Things to consider

Talk about the different aspects about network viz the user needs to consider: layout, vertex size, vertex colour, vertex shape, edges, edges width, etc. Talk about the different components and how can we use them (to represent what, for example.) The size of the network, and type of the network (egocentric, small, large, bipartite, etc.)

In terms of the layouts, what are the things we need to consider (we can mention R packages that implement layouts in R).

Start to finish example

Example 1: School data

Cleaning Data

First, the data needs to be pulled in. It can be found at (INSERT LINK HERE). After we pull it in, let’s get a glimpse of what the data looks like.

# attaching packages
library(igraph)
library(data.table)
library(devtools)
install_github("USCCANA/netplot")
library(netplot)


# loading and cleaning data
students <- fread("./data/middle_school/pone.0153690.s001.csv")
interactions <- fread("./data/middle_school/pone.0153690.s003.csv")

print(students)

##        id grade gender unique lunch initialsNum
##   1: 2003     7      0      0     1         386
##   2: 2004     8      1      1     1         402
##   3: 2006     7      1      1     2         288
##   4: 2008     8      0      1     1         199
##   5: 2009     7      1      0     1         147
##  ---                                           
## 674:   NA     8      0      0    99         171
## 675:   NA     8      0      1    99         270
## 676:   NA     8      0      1    99         327
## 677:   NA    99      1      0    99         378
## 678:   NA     7      1      0    99         277

print(interactions)

##          id contactGender contactGrade contactId ClassPeriod contactInitialNum
##     1: 2004             1            8      3127           4               323
##     2: 2004             0            8      2620           1               335
##     3: 2004             1            8        99           1               401
##     4: 2004             1            8        99           9               401
##     5: 2004             1            8        99           9               401
##    ---                                                                        
## 10777: 3448             1            7        99           4                79
## 10778: 3448             1            7        99           2                17
## 10779: 3448             1            7        99           4                17
## 10780: 3448             1            7      3439           3               155
## 10781: 3448             1            7        99           3               294

In order to use the data, we need to remove all of the ’N/A’s and miscoding in the datasets. Also, we see a large number of students who only have interactions with themselves (they do not interact with anyone else through the day), so these “isolates” need to be removed in order for the graph to be more easily read.

# filtering out 'N/A's in the 'students' data frame
students <- students[!is.na(id)]

# filtering down to gender being "0" or "1"
students <- students[gender %in% c("0", "1")]

# filter out 'N/A's in 'id' and 'contactId' 
interactions <- interactions[!is.na(id) & !is.na(contactId)]

# Which connections are not OK?
ids <- sort(unique(students$id))

# narrowed our data from 10781 to 5150
interactions <- interactions[(id %in% ids) & (contactId %in% ids)]

source(file = "./misc/color_nodes_function.R")

After, the two datasets need to be combined together.

## Creating matrix from datasets
net <- graph_from_data_frame(
  d = interactions[, .(id, contactId)],
  directed = FALSE, vertices = as.data.frame(students)
)

## Getting only connected individuals
net_with_no_isolates <- induced_subgraph(net, which(degree(net) > 0))

Finally, we plot it, effectively showing this network graph.

## Plot with no isolates
nplot(
  net_with_no_isolates
) 

Split according to Grade

Here, we are taking the data set and the plot, letting us customize a number of aspects of the graph. First, in order to work with the “color_nodes” function, we need to make “grade” a factor instead of being numeric. Also, we identify the colors we would like the nodes to be.

  ## adjust 'grade' to factor 
V(net_with_no_isolates)$grade <-  as.factor(V(net_with_no_isolates)$grade)  

# plotting connections among grades ####
set.seed(77) 

a_colors <- color_nodes(net_with_no_isolates,"grade", c("gray40","red3"))
attr(a_colors, "map")

##         7         8 
## "#666666" "#CD0000"

Now, we are able to create a plot of the data. This is the same data that we used to create the plot above, but now adjustments to the nodes will be made. - Color the vertices (‘vertex.color’) according to the grade the student is in (with 7th graders being gray and 8th graders being red). - Adjust the shape of the vertices (‘vertex.nsides’). If the student is a 7th grader, the vertices will be a circle, but if they are not, the vertices will be a triangle. - Adjust size of vertices (‘vertex.size.range’).

grades <- nplot(
  net_with_no_isolates,
  vertex.color = color_nodes(net_with_no_isolates, "grade", c("gray40","red3")),
  vertex.nsides = ifelse(V(net_with_no_isolates)$grade == 7, 10, 3),
  vertex.size.range = c(0.015, 0.2))

grades <- nplot(
  net_with_no_isolates,
  vertex.color = color_nodes(net_with_no_isolates, "grade", c("gray40","red3")),
  vertex.nsides = ifelse(V(net_with_no_isolates)$grade == 7, 10, 3),
  vertex.size.range = c(0.015, 0.2),
  edge.color = ~ego(alpha = 1, col = "lightgray") + alter(alpha = 0.25, col = "lightgray"),
  vertex.label = NULL,
  edge.curvature = pi/6,
  edge.line.breaks = 10  
)  

# add radial gradient fill 
grades <- set_vertex_gpar(grades, 
                          element = "core",
                          fill = lapply(get_vertex_gpar(grades, "frame", "col")$col, \(i) {
                            radialGradient(c("white", i), cx1=.8, cy1=.8, r1=0)
                          }))

# add legend to graph
grades_general <- nplot_legend(
  grades,
  labels = c("7th", "8th"),
  pch = c(21,21),
  gp = gpar(
    fill = c("gray40","red3")),
  packgrob.args = list(side = "bottom"),
  ncol = 2  
)


grades_general
grid.text("Split According to Grade", x = .2, y = .87, just = "bottom") 

Split according to Gender

Next, the data will be split according to gender, with male being yellow and female being green. Also, the male points are circles, while the female points are diamonds.

# let's get a graph for the gender data
V(net_with_no_isolates)$gender <-  as.factor(V(net_with_no_isolates)$gender)

a_colors <- color_nodes(net_with_no_isolates,"gender", c("lightgoldenrod2","forestgreen"))
attr(a_colors, "map")

##         0         1 
## "#EEDC82" "#228B22"

## plot
set.seed(77)
gender <- nplot(
  net_with_no_isolates,
  vertex.color = color_nodes(net_with_no_isolates, "gender",c("lightgoldenrod2","forestgreen")),
  vertex.nsides = ifelse(V(net_with_no_isolates)$gender == 0, 10, 4),
  vertex.size.range = c(0.01, 0.01),
  edge.color = ~ego(alpha = 0.33, col = "gray") + alter(alpha = 0.33, col = "gray"),
  vertex.label = NULL,
  edge.line.breaks = 10
)

# add legend to graph
nplot_legend(
  gender,
  labels = c("Male", "Female"),
  pch = c(21,23),
  gp = gpar(
    fill = c("lightgoldenrod2","forestgreen")),
  packgrob.args = list(side = "bottom"),
  ncol = 2
)

grid.text("Split According to Gender", x = .2, y = .87, just = "bottom") 

Split according to Lunch Period

Here is the data split according to the different lunch periods students might be in.

# now let's do the same with lunch period 
V(net_with_no_isolates)$lunch <-  as.factor(V(net_with_no_isolates)$lunch)

a_colors <- color_nodes(net_with_no_isolates,"lunch", c("purple","palegreen","steelblue"))
attr(a_colors, "map")

##         1         2        99 
## "#A020F0" "#98FB98" "#4682B4"

## plot
set.seed(77)
lunch <- nplot(
  net_with_no_isolates,
  vertex.color = color_nodes(net_with_no_isolates, "lunch",c("purple","palegreen","steelblue")),
  vertex.nsides = 
  ifelse(V(net_with_no_isolates)$gender == 0, 4,   # First Lunch
        ifelse(V(net_with_no_isolates)$gender == 1, 3,  # Second Lunch 
               10)),                                 # Other
  vertex.size.range = c(0.01, 0.01),
  edge.color = ~ego(alpha = 0.33, col = "gray") + alter(alpha = 0.33, col = "gray"),
  vertex.label = NULL,
  edge.line.breaks = 10
)

# add legend to graph
nplot_legend(
  lunch,
  labels = c("First", "Second", "Other"),
  pch = c(23,24,21),
  gp = gpar(
    fill = c("purple","palegreen","steelblue")),
  packgrob.args = list(side = "bottom"),
  ncol = 3
)

grid.text("Split According to Lunch Period", x = .2, y = .87, just = "bottom") 

Changing Lines to Dashes

One of the perks of netplot is the ability to be fully customizable, right out of the box. First, here is an example of the same dataset, but the edges are dashed instead of full and straight instead of curved.

set.seed(77)

grades <- nplot(
  net_with_no_isolates,
  bg.col = "#F5F5F5",
  vertex.color = color_nodes(net_with_no_isolates, "grade", c("red","blue")),
  vertex.size.range = c(0.02, 0.02),
  edge.color = ~ego(alpha = .15, col = "black") + alter(alpha = .15, col = "black"),
  vertex.label = NULL,
  edge.width.range = c(2,2),
  edge.line.lty = 6,
  edge.line.breaks = 1  
)  



# add legend to graph
grades_dashed <- nplot_legend(
  grades,
  labels = c("7th", "8th"),
  pch = c(21,21),
  gp = gpar(
    fill = c("red","blue")),
  packgrob.args = list(side = "bottom"),
  ncol = 2  
)

grades_dashed

Colored Edges and Skipped Vertices

This selection shows how to skip vertices and add colors to the edges.

set.seed(77)   

grades <- nplot(
  net_with_no_isolates,
  bg.col = "#F5F5F5",
  vertex.color = color_nodes(net_with_no_isolates, "grade", c("red","blue")),
  vertex.nsides = ifelse(V(net_with_no_isolates)$grade == 7, 10, 4),
  vertex.size.range = c(0.0001, 0.0001),
  edge.color = ~ego(alpha = 0.33) + alter(alpha = 0.33),
  vertex.label = NULL,
  edge.width.range = c(2,2),
  edge.line.breaks = 10
  )  

# add legend to graph
grades_edge_colored <- nplot_legend(
  grades,
  labels = c("7th", "8th"),
  pch = c(21,21),
  gp = gpar(
    fill = c("red","blue")),
  packgrob.args = list(side = "bottom"),
  ncol = 2
  )

grades_edge_colored

Changing Background Color

We can also add a background to the plot, including a gradient.

set.seed(77)   

grades <- nplot(
  net_with_no_isolates,
  bg.col = linearGradient(c("lightpink", "lightskyblue")),
  vertex.color = color_nodes(net_with_no_isolates, "grade", c("red","blue")),
  vertex.nsides = ifelse(V(net_with_no_isolates)$grade == 7, 10, 4),
  vertex.size.range = c(0.01, 0.01),
  edge.color = ~ego(alpha = 0.15, col = "black") + alter(alpha = 0.15, col = "black"),
  vertex.label = NULL,
  edge.line.breaks = 10  
)  

# add legend to graph
grades_background <- nplot_legend(
  grades,
  labels = c("7th", "8th"),
  pch = c(21,23),
  gp = gpar(
    fill = c("red","blue")),
  packgrob.args = list(side = "bottom"),
  ncol = 2  
)

grades_background

Different Colors

We can also change vertex colors and edges to straight lines.

set.seed(77)

grades <- nplot(
  net_with_no_isolates,
  bg.col = "#F5F5F5",
  vertex.color = color_nodes(net_with_no_isolates, "grade", c("#FFDB58","#708090")),
  vertex.nsides = ifelse(V(net_with_no_isolates)$grade == 7, 10, 4),
  vertex.size.range = c(0.02, 0.02),
  edge.color = ~ego(alpha = .15, col = "black") + alter(alpha = .15, col = "black"),
  vertex.label = NULL,
  edge.width.range = c(2,2),
  edge.line.breaks = 1
)

# add legend to graph
grades_different_color <- nplot_legend(
  grades,
  labels = c("7th", "8th"),
  pch = c(21,23),
  gp = gpar(
    fill = c("#FFDB58","#708090")),
  packgrob.args = list(side = "bottom"),
  ncol = 2  
)

grades_different_color

Example 2: LTCF data

# loading packages
library(network)
library(netplot)  

# loading data 
load("./data/nursing_home/network99_f1.RData")  

Create for-loop:

# Creates an empty list to store the networks
nets <- list()  

# Sets a seed for reproducibility 
set.seed(1231)  

for (i in 1:99) { 
  # Checks if the vertex "is_actor" exists in the network
  is_health_care_provider <- networks[[i]] %v% "is_actor"  
  nets[[i]] <- nplot( networks[[i]], 
                      # Colors the vertices gray if HCP exists, red otherwise
                      vertex.color     = ifelse(is_health_care_provider, "gray40", "red3"),  
                      # Makes vertices square if HCP exists, round otherwise
                      vertex.nsides    = ifelse(is_health_care_provider == TRUE, 4, 10),
                      # Makes HCP vertices larger than patient vertices
                      vertex.size      = ifelse(is_health_care_provider == TRUE, .25, .15),
                      vertex.size.range = c(.015,.065),
                      edge.width.range = c(.25,.5),
                      # Sets edge line breaks to 1 and colors edges black  
                      edge.line.breaks = 1, edge.color = ~ ego(alpha = 1, col = "lightgray") +    alter(alpha = 1, col = "lightgray"),
                      edge.curvature = pi/6,
                      # Removes vertex labels
                      vertex.label     = NULL )
}

# Combines the 99 plots into an 11x9 grid
allgraphs <- gridExtra::grid.arrange(grobs = nets, nrow=11, ncol=9)